NAFO Sci. Coun. Studies, 18: 81–85 The Greenland ( morhua) at Iceland 1941Ð90 and Their Impact on Assessments

Sigfús A. Schopka Marine Research Institute, Skúlagata 4 121 Reykjavik, Iceland

Abstract Studies have shown cod (Gadus morhua L.) in Icelandic waters form a unit stock with very little emigration, while cod from East and West Greenland migrate to Iceland and affect the stock abundance there. Such immigrations have through recent history resulted in the overestimating of the Iceland stock size while underestimating fishing mortality. Noting that for Icelandic cod the catch-at-age peaks at ages 4 or 5 years while immigrations from Greenland causes it to peak again at 7–9 years, Virtual Population Analysis was carried out on Icelandic stock analysis data from the period 1941–90 to determine years where fishing mortalities were higher for the older age groups. Immigrating year-classes through the period were thereby identified, particularly the outstanding 1945 year-class. On the other hand, it was observed that some year-classes emigrating from Greenland had in fact originated from spawning grounds at Iceland and drifted as egg/larvae to Greenland. This demonstrated the importance of information on larval drift.

Key words: Cod, Gadus morhua, exploitation pattern, fishing mortality, Iceland–Greenland, migrational rate, stock assessment

Introduction mortalities for those year-classes at Iceland which are subject to such an immigration. Tagging experiments at Iceland show that cod (Gadus morhua L.) in Icelandic waters are more or Materials and Methods less a unit stock (Jónsson, 1990). From more than 100 000 cod tagged since 1948, less than 30 recap- For this analysis, data on catch-in-numbers were tures have been recorded outside the Icelandic available dating back to 1941. Data for the most fishing grounds (J. Jónsson, pers. comm.). On the recent years, i.e. 1971 and later, were basically the other hand, it is well known that cod from Greenland same as in Anon. (1991) with some modifications waters, both West and East Greenland stocks, do for cod older than 14 years. For the period 1955–70 migrate in some years even on a large scale to basic data were the same as in Anon. (MS 1976) but Iceland (Hovgård et al., MS 1989; Harden Jones, data on the Icelandic fisheries were revised based 1968). These immigrations affect the stock abun- on detailed information on how catches were dis- dance at Iceland and hence the management of the tributed by gear and area. Also, mean weights-at- stock. It is therefore important, in order to under- age have been revised to obtain more reliable SOP- stand the fluctuations in the stock at Iceland, to be figures than those which can be calculated from the able to distinguish between the two cod stock com- data in Anon. (1976). The SOP-values between ponents, i.e. the Icelandic and the Greenlandic 1955 and 1970 give much higher landings than in components. reality. Therefore, mean lengths-at-age available for different fisheries were converted to mean In general there is no exact solution available to weights-at-age and then weighted by the catch of the problem at present. Theoretically tagging ex- each fishery to give the final figures of the total periments would solve this under the assumption fisheries. The revised mean weights were then used that all recaptures were available for analysis and for the stock biomass calculations in the above- that tagging mortality, fishing mortality and natural mentioned period. Estimated catch-in-numbers mortality for both components are known. As these for the period 1941–52 were entirely based on assumptions are never met, analysis of tagging Icelandic age-length data available at MRI but in experiments have only provided average estimates some cases never been published or worked up (Anon., MS 1981). Other parameters which are until recently (Schopka in preparation). Prior to different for these stocks, such as otolith structure 1953, only length composition data were available (Rätz, MS 1990), growth rate etc., have not yet for the German fisheries but age-composition data answered this question sufficiently. were also available for the period 1953–54. By using the Icelandic age-length keys, the German In this paper an attempt is made to analyze the length composition data were used to account for immigration by studying the changes in the fishing that fishery. 82 Sci. Council Studies, No. 18, 1993

Another important cod fishery at Iceland in those TABLE 1. Cod at Iceland: F-at-age from an ordinary years was that of the United Kingdom (UK). The VPA run for the 1960 to 1964 year-classes. only available information on that fishery prior to 1955 were catch and effort data by fishing grounds Year-class and by time of year. Icelandic age-composition Age 1960 1961 1962 1963 1964 data for the same time and grounds were then used to convert the UK landings into catch-in-numbers 3 0.12 0.08 0.10 0.08 0.09 data. 4 0.28 0.17 0.20 0.16 0.25 5 0.41 0.16 0.24 0.19 0.40 6 0.42 0.13 0.25 0.22 0.40 For the at Iceland catch-by-age 7 0.52 0.27 0.31 0.30 0.52 normally reaches a peak at age groups 4 or 5 years. 8 1.08 0.44 0.57 0.38 0.60 For a typical Greenland year-class entering Iceland 9 0.76 1.05 0.60 0.70 0.68 grounds, catch in numbers peaks to fish at age 7 or 10 0.69 1.11 1.03 1.11 1.04 8 years, or even 9 years in the case of a strong year- class such as the 1945 year-class (Fig.1). From the available catch-in-numbers data an ordinary Virtual TABLE 2. Cod at Iceland: F-at-age from a VPA run Population Analysis (VPA) was run. Input fishing where F-values have been in- creased mortality (F) for the oldest age group was chosen as to account for the im-migration of 1961– 63 year-classes. 0.5 but terminal Fs in 1990 were the same as in Anon. (1991). It turned out that, as expected for Year-class some year-classes, Fs for the younger age groups were extremely low. Based on the assumption that Age 1960 1961 1962 1963 1964 the exploitation pattern had not changed, it was 3 0.12 0.11 0.12 0.10 0.09 expected that Fs for these age groups should not be 4 0.28 0.25 0.24 0.23 0.25 lower than for the surrounding year-classes at same 5 0.41 0.27 0.31 0.30 0.40 age and the difference can be explained as an 6 0.42 0.24 0.35 0.40 0.40 effect of an immigration. To account for this, F on 7 0.52 0.60 0.50 0.50 0.52 these age groups can be increased. In the present 8 1.08 0.70 0.70 0.70 0.60 paper, F was increased to a level of a similar order 9 0.76 1.05 0.80 0.70 0.68 as that of the surrounding year-classes at the same 10 0.69 1.11 1.03 1.11 1.04 age, or to a level representing the likely fishing pattern. An example of the method for the 1960–64 year-classes is shown in Tables 1 and 2. It should data or effort data are available, F values can be be noted that the underlined F-values in Table 2 selected by tuning (Gunnar Stefánsson, Marine Re- were chosen as input values. If groundfish survey search Institute, Iceland, pers. comm.). In the case of Icelandic cod, such data were only available for the most recent years. It can of course be argued that the F-values chosen here are not the right ones, but it is the author’s opinion that they can be seen as fairly realistic. 40 Results

30 Applying the revised fishery mortalities led to the identification of the following year-classes as having a high number of immigrants from Greenland: 1936, 1937, 1938, 1942, 1945, 1950, 1953, 1956, 20 1961, 1962, 1963, 1973 and 1984 (Fig. 2). It is quite possible that some cod from other year-classes

Catch in numbers (millions) 1945 might well have also migrated to Iceland, but if they 1944 existed those migrations would have been so small 10 that they could not be detected from the usual noise 1946 in the data.

0 Most abundant of all the Greenland year-classes 3456 789101112 Age (years) in the period in question was the outstanding 1945 year-class. According to landings from this year- Fig. 1. Cod at Iceland: catch-at-age for the 1944–46 class in the late-1940s and the beginning of the year-classes. 1950s, this year-class turned out to be of an aver- SCHOPKA: Greenland Cod at Iceland and Their Impact on Assessments 83

700 Icelandic cod Immigrants 600

500

400

300 Numbers (millions) 200

100

0 1938 40 45 50 55 60 65 70 75 80 85 Year-classes

Fig. 2. Cod at Iceland: year-class size at age 3. age size at Iceland. However, suddenly in 1953 and years around 1950 when the actual stock biomass the following years, this year-class showed up in at Iceland was overestimated by 40-60%. When the large quantities on the spawning grounds off the 1973 and 1984 year–classes immigrated to Iceland, southwestern coast of Iceland (Fig. 1). According the stock biomass was overestimated by 15-20% to the present analysis, two thirds of this year-class (Fig. 3). on Iceland grounds were immigrants from Greenland (Fig. 2). For other Greenland year-classes, around Even though the fishing mortality at younger one third or less of the year-class size at Iceland ages of an immigrating year-class is lower than for were immigrants from Greenland. the surrounding year-classes, the average unweighted fishing mortality for the most exploited In years prior to 1970 when Greenland year- age groups does not differ much when comparing classes were more frequent at Iceland, one or more the Icelandic stock component to the combined Greenland year-classes could be found at the same estimates (Fig. 4). On the other hand, if fishing time in the stock at Iceland every year. In such mortality is weighted by the year-class strength, a cases the VPA overestimates the stock size at Ice- much larger difference appears compared to land considerably, as the VPA back-calculates the unweighted fishing mortality depending on the mag- immigrants and Icelandic cod as one single stock nitude of the immigration and the relative year-class and also includes the immature young cod compo- strength at Greenland and Iceland. This was espe- nent still at Greenland as a part of the Icelandic cially pronounced in early-1950s. immature stock at Iceland (Fig. 3).

The impact of the immigrants on the assessment is Discussion greater than reflected by year-class strength only. This is due to the fact that the immigrants are almost entirely According to the analysis presented here, im- large mature cod, which have more impact on the migrations took place more often prior to 1970 than biomass estimate of the stock than on the stock size in during the last two decades. This might be con- numbers. Thus, for example, let us look again on the nected to a much poorer state of the stock at 1945 year-class. If no fish of the 1945 year-class had Greenland especially West Greenland after 1970. immigrated in 1953, the expected biomass of 8-year- On the other hand, the last three year-classes which old fish would have been around 200 000 tons. The emigrated from Greenland, i.e. 1963, 1973 and large scale immigration changed that figure to more 1984 year-classes, were all known to be of Icelandic than 900 000 tons of 8-year-old fish. origin to some extent. They were described to have drifted as eggs/larvae from the spawning grounds The migration of an immigrating year-class is at Iceland (Anon. 1976, Vilhjálmsson and Magnússon normally spread over 2–3 years and therefore more 1984). Due to lack of research information prior to than one immigrating year-class can be present at 1963, no observation have been reported on larval the same time in the stock at Iceland. This further drift from Iceland. However, it may be concluded increases the discrepancy in the back–calculations that since the three above-mentioned year-classes of the stock. The difference was greatest in the were of Icelandic origin, the other previous immi- 84 Sci. Council Studies, No. 18, 1993

3

2

1 Biomass (million tons)

0 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990

Fig. 3. Cod at Iceland: stock biomass estimates. Broken line: ordinary VPA run; solid line: VPA run based on expected or likely fishing mortality at Iceland. This line should reflect the stock biomass at Iceland each year.

1.0 A 0.8

0.6

0.4

0.2

0 B 0.6

0.4 Fishing mortality of age group 5–10 0.2

0 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 Fig. 4. Cod at Iceland: (A) unweighted mean fishing mortality for the age groups 5–10; (B) weighted mean fishing mortality for the age groups 5–10. Broken lines: ordinary VPA run; solid lines: VPA run based on expected or likely fishing mortalities at Iceland. grating year-classes could also have originated From the year-class strength shown in Fig. 2, from the spawning at Iceland. If this is the case, the the size of the year-class at age 3 has fluctuated by decline in frequency of larval drift towards Greenland a factor of 7. It should, however, be noted that the may be caused by changes in the environment and/ year-class strength in Fig. 2 only reflects the num- or be connected with the reduction of the stock at bers in the stock at Iceland. A part of the year- Iceland. classes which drifted to Greenland were exploited SCHOPKA: Greenland Cod at Iceland and Their Impact on Assessments 85 there and therefore the actual size of these year- biomass and lower fishing mortality, especially on classes are underestimated. If data are available the younger age-groups. Total allowable catch for the Greenland fisheries, it might be possible to (TAC) recommendations for 1982 of 450 000 tons shed some light on the actual size of these year- was based on that assessment. Even though cod classes. When studying the stock recruitment rela- catches were somewhat restricted at that time (ef- tionship for the cod stock at Iceland or fluctuations fort limitations), the stock was in such bad shape in the recruitment in relation to the environment, that the total catch (388 000 tons in 1982) did not such information is important. reach the recommended TAC.

Even though calculation from tagging experi- It is important both for the assessment and ments at Greenland indicate an average emigration management of the Icelandic cod fisheries, that rate of 0.5 at West Greenland and a value of 0.29 at information on larval drift and the size of the stock East Greenland (Anon., MS 1981), the emigration components is collected. This can be done by fluctuates considerably between years depending carrying out more extensive 0-group and ground- on year-class strength, age-at-maturity, etc. Other fish surveys in the Iceland-Greenland areas. year-classes than those defined here as immigrat- ing year-classes might also have included some Acknowledgements migrants, but in comparison to the stock size at Iceland, two or three million immigrants are hardly The author thanks the following scientists for detectable and do not affect the assessment at all. their help in providing data on the United Kingdom and German fisheries at Iceland during the years More serious consequences for the stock as- 1946–54: A. C. Burd, D. Cushing and B. W. Jones, sessment at Iceland could happen when a strong Lowestoft, U.K. and the late A. Meyer, Hamburg, immigrating year-class does show up on the Icelan- Germany. dic grounds. This was not a problem earlier when fishing was unlimited and VPA was unknown. At that References time everybody, including the fishery scientists, was happy if more fish were seen on the fishing ANON. MS 1976. Report of the North-Western Working grounds. After the stock declined and the fishery Group. ICES C.M. Doc., No. F:6. needed to be restricted, immigrants posed an extra MS 1981. Report of the Working Group on cod stocks off East Greenland. ICES C.M. Doc., No. G:6. problem for the assessment of the stocks and hence MS 1991. State of marine stocks and environ- the management. In 1980 and 1981, the 1973 year- mental conditions in Icelandic waters 1991. Pros- class showed up in large quantities on the spawning pects for the fishing year 1991/92. grounds off the southwest coast of Iceland. Back- Hafrannsóknastofnunin. Fjölrit nr. 25, Reykjavik, 1991, calculations by VPA showed an increase in the 153 p. (In Icelandic). stock size of the younger ages of the year-class, HARDEN JONES, F. R. 1968. Fish migration. Edward and hence the fishing mortality in the years 1977–79 Arnold, London, England. decreased in the new analysis compared to the HOVGÅRD, H., F. RIGET, and H. LASSEN. MS 1989. Modelling cod migration from Greenland to Iceland. assessment made the year before. NAFO SCR Doc., No. 32, Serial No. N1608, 18 p. JÓNSSON, J. 1990. Hafrannsóknir vid Ísland eftir 1937. As this was shortly after the extension of Icelan- Vol. 2, 447 p. Menningarsjódur. (Marine research in dic fisheries jurisdiction which led to the sharp Icelandic waters after 1937). (In Icelandic) reduction of non-Icelandic fishing effort in 1977 and RÄTZ, H. J. MS 1990. The assessment of the migration of 1978, this decline in fishing mortality (weighted Fs (Gadus morhua L.) between the stocks was used at that time) was consequently inter- off East and West Greenland by means of otolith preted as a part of the overall reduction of the typing. NAFO SCR Doc., No. 34, Serial No. N1751, 42 p. fishing effort, and the impact of the immigration VILHJÁLMSSON, H. and J. V. MAGNÚSSON. MS 1984. which could not be quantified very well at that time Report of the 0-group fish survey in Icelandic and was underestimated. The cod stock at Iceland was East Greenland waters, August 1984. ICES C.M. Doc., therefore thought to be in a better state, i.e. larger No. H:66. Blank